Implementation of genomic medicine for gastrointestinal tumors

Abstract Genomic medicine is an approach to take advantage of genomic data in medical practice and health care. The advancement of sequencing technologies has enabled the determination of individual genomes as well as the genome in neoplasms. In the field of human cancer, understanding genomic alterations in tumors and variations associated with drug responses has paved the way towards the development of new drugs and personalized medicine. International collaborations of cancer genome analyses have accumulated a huge body of information about somatic mutations, and identified new driver mutations and pathways in a wide range of cancers. In particular, a growing body of evidence has shown that information about mutations in neoplasms helps to assess the efficacy and resistance of anti‐cancer drugs. Information about germline mutations associated with hereditary cancer has been shown to benefit patients by enabling early detection of their tumors and disease‐specific treatment, as well as reducing the risk for those at risk. To promote personalized medicine in a more cost‐effective and personalized way, further inter‐institutional, nationwide, and international collaboration is needed. This article summarizes the background and current situation of genomic medicine in the field of gastrointestinal tumors to help physicians and medical coworkers by assisting their better understanding of genomic medicine and strengthening their confidence of its clinical use.

Implementation of genomic medicine has emerged in many developed countries. In the USA, several companies and institutions launched services and projects to use somatic mutation data in personalized cancer treatment. Application of genomic medicine in clinics has been supported by the US government, as the former president, President Obama, launched the precision medicine initiative in January 2015. In 2016, the French government decided to invest 670 million euro in promoting genomic medicine and personalized medicine. Genomics England started whole genome sequencing of 100 000 patients with rare diseases and their families, as well as patients with common cancers. Although Japan was a slow starter, the Japanese government decided to support the development of systems and infrastructures for personalized medicine in the field of rare undiagnosed diseases and cancer. It is obvious that collaboration is of paramount importance for the implementation of personalized medicine. Adequate knowledge and strategies of genomic medicine need to be shared not only by the researchers and informaticians in medical and genomic studies, but also by physicians and medical coworkers in clinics, and together they need to make efforts to benefit cancer patients in a cost-effective way ( Figure 1).
The aim of this article is to overview four critical issues for the implementation of genomic medicine in gastroenterological oncology, namely: (i) personalized cancer treatment; (ii) incidental/secondary findings; (iii) assessment of cancer susceptibility; and (iv) adverse effects and resistance of anti-cancer drugs.

PE RSONALIZED CANCER TREATMEN T
Anti-cancer drugs have dramatically changed over the years, from drugs suppressing universal mechanisms of cell growth to those specifically targeting cancer-driver molecules or pathways. Identification of chimeric proteins in neoplasms led to the development of a new generation of drugs such as imatinib and dasatinib for BCR-ABL fusion proteins, and crizotinib and alectinib for ALK fusions. In addition, several receptor tyrosine kinases have become new targets of anti-cancer drugs because their kinase activity controls the growth of some neoplastic cells. For example, gefitinib was developed to inhibit the kinase activity of epidermal growth factor receptor (EGFR) and used for the treatment of non-small-cell lung cancer. Later, genetic analysis disclosed that the efficacy of gefitinib was closely linked with the activating mutations in its cytoplasmic region. Now, analysis of genetic alterations in the cytoplasmic region is mandatory as a companion diagnosis for the use of EGFR inhibitors including gefitinib, erlotinib, and afatinib. It is of note that cancer cells are "addicted" to the enhanced growth signaling by the mutations in EGFR. However, therapeutic antibodies that antagonize EGFR such as cetuximab and panitumumab are used for patients with wild-type EGFR. These antibodies do not benefit patients carrying a RAS mutation in their tumors because the downstream signal of EGFR is activated by KRAS or NRAS mutation. Therefore, screening of RAS mutations is essential for the exclusion of patients that will not benefit from the drugs. Moleculartargeted drugs for gastrointestinal cancer, their targets, and companion diagnostics (CDx) are summarized in Table 1.
Development of immune checkpoint inhibitors, which are new anti-cancer drugs that suppress deregulated immune checkpoint systems, has rapidly led to their application to various types of solid tumors including malignant melanoma, non-small-cell lung cancer, colorectal cancer, and other tumors. As these drugs are reactivating the immune system against cancer by hampering the immune-escaping mechanism(s) of cancer cells, one of the factors affecting their efficacy is assumed to depend on the number of antigens expressed on the surface of tumor cells. Consistent with this view, a clinical study showed that pembrolizumab, an anti-programmed cell death 1 (anti-PD-1) anti- biomarker for immune checkpoint inhibitors. 2 Challenges are underway to discover additional candidate biomarkers.
Next-generation sequencing has been adopted in many clinical laboratories because of its low cost and rapid speed of sequencing compared with the traditional Sanger's method. As the analysis of target molecules and/or their downstream signals leads to the discovery of predictive markers for the treatment of several anti-cancer drugs, analysis of multiple genes is now considered to be an effective strategy to select drugs. A wide range of cancer panels such as hot-spot mutation panels, actionable gene panels, and comprehensive gene panels have come into clinical practice. A population-based study using a panel with 212 amplicons of 48 genes showed that clinically relevant mutations or mutations associated with human carcinogenesis were found in approximately 63% (534/854) of patients with a variety of cancers, and that actionable mutations or mutations providing information of the sensitivity or resistance to approved and preclinically available drugs were identified in about 26% of patients. 3 Another study using a panel containing 189 amplicons of 46 genes showed that clinically relevant mutations were found in approximately 87% (296/342) of patients with mainly melanoma, non-small-cell lung cancer, and colorectal cancer, and that actionable mutations were identified in about 35% (122/351) of patients. 4 Although the frequencies of mutations  11 Although the physicians enrolled in the study indicated the presence of an actionable mutation in 805 of 1474 cases, expert curators judged 362 of 805 cases as actionable, suggesting that involvement of experts is crucial for the appropriate implementation of genomic medicine. In addition, clinical actionability of pathogenic variants is defined by evidence for their potential utility as therapeutic targets, but real actionability may be different depending on the country, health-care system, insurance, accessibility to drugs, and/or various situations of individuals.

IN NGS PANEL TESTS FOR SOMATIC MUTATIONS
As panel sequencing includes genes responsible for cancer susceptibility, it is possible to identify germline mutations in cancer tissues.
In the case of analysis of tumor tissue alone, the origin of mutations may be indistinguishable from germline variants, but other clinical

SUSCEPTIBILITY FOR EVIDENCE-BASED HEALTH CARE
Prevention and early detection are two of the most effective strategies to avoid death from malignant neoplasms. Family history is important information for the assessment but the determination of deleterious mutation in the responsible genes will give us more ben-

| AD VERSE EFFE CTS AND RESISTANCE OF ANTI-CANCER DRUGS
Adverse effects are one of the major problems that hamper the use These pieces of information are useful to understand the mechanisms of resistance in refractory tumors, and may be applicable for selection of other drugs or modalities that are effective in secondary mutation. However, in many cases, resistance originates from multiple non-mutational, non-genetic mechanisms. 30 Therefore, incorporation of transcriptome, epigenome, proteome, and metabolome into genomic information is needed for precise assessment of resistant mechanism(s). Integration of these data together with information on drug efficacy, toxicity, and resistance will be a vital challenge for the promising development of specific and effective precision medicine. 31

| CHALLENGES FOR BETTE R PRECISION ONCOLOGY
Although cost-effectiveness of genetic testing is an important matter of concern from the financial point of view, it will not be discussed here because it is beyond the scope of this review. Current challenges for medical oncology include the interpretation of variations in the individual genome and in the cancer genome. Larger collaborations and data repositories are needed for the assessment of less common variants, which otherwise might elude statistical analysis.
Accumulating genomic data are rapidly expanding and now overflowing beyond our recognition. In the near future, we will obtain epigenomic, proteomic, metabolomic, and microbiota information that will require clinical interpretation. Therefore, it is a matter of pressing necessity to develop integrative databases and analytical systems for tumor DNA will widen the approaches for precision medicine. More importantly, the development of applicable drugs for a wide range of driver alterations is essential for the promotion of genomic medicine.
The number of current molecular targeted drugs for gastrointestinal tumors is limited compared with those for hematological malignancies and lung cancer.
In conclusion, because information and technology are rapidly expanding in the field of genomic analyses, gastroenterologists need to collaborate with oncologists, geneticists, pharmacologists, computational biologists, and bioinformaticians for the implementation of genomic medicine in gastrointestinal malignancies.

DISCLOSURE
Conflicts of Interest: Author declares no conflicts of interest for this article.